Technical considerations for saving lives

Clean water is an engineering problem and a community problem at the same time. Our technical team works out the science, and each local partner knows what their community will actually use. We iterate between the two, often five times over, until a solution is scientifically sound, wanted by the community, and built to last. This is where we show that work.

We don’t just apply academic science and engineering. We add systems thinking and economics, and we design closely with local NGOs so every solution is practical. That means iterating with the community, and usually starting with a pilot, until it works in the real world.

However deep you want to go, there’s a path

What does clean water actually cost?

We track impact-per-dollar closely. Here’s the approximate cost to give one person safe water for a year: the initial delivered cost spread over each system’s expected life. Routine upkeep, paid locally, is not included. Judge for yourself.

Swipe the table sideways to see every column →

SolutionLocationCost (estimate)People servedExpected life$ / person / yrEst. cost per life saved
Shared gravity membrane filterDR Congo (Goma)~$80 per filter, all-in~25 (5 households)~5 years~$0.64~$1,000
Pipelines + filtration at the tapsPapua New Guinea~$25,000 for the full plan~8,500~10 years~$0.30~$450
School ultrafiltration systemCambodia (Kampot)~$700 per school~200–300 students~10 years~$0.25–0.35~$500
Ultrafiltration + UV village systemPhilippines (Mariveles)$533~100–200 a day~5 years~$0.55–1.05Not estimated
Household biosand filterCambodia (Siem Reap)~$100~7–2010+ years~$0.50–1.40Not estimated

$ per person per year is the initial delivered cost of each system spread over its expected life and the people it serves. Routine upkeep is paid locally and is not included. DR Congo: $80 per filter covers the filter, delivery, installation, training and materials. Papua New Guinea: about $5,000 for filtration on the existing line (3,500+ people whose water is untreated today), about $5,000 to extend the pipeline to about 1,500 more people, and about $15,000 for a new pipeline for about 3,500 more people. These are estimates; filtration and construction costs vary and depend on things we don’t know yet. Cambodia: the range reflects 200 to 300 students per school.

How we estimate cost per life saved

This is an estimate, not a measured result. We use it only for DR Congo, Papua New Guinea and Cambodia, where the people we serve have no safe water source today.

  1. People without safe water. About 703 million people lack even a basic drinking water service: 292 million walk more than 30 minutes to an improved source, 296 million drink from unprotected wells and springs, and 115 million drink untreated surface water (WHO and UNICEF, 2022 data). We use this group, not the 2.2 billion who lack safely managed water, because that larger number includes many people who already have a protected source nearby.
  2. Deaths from unsafe drinking water. Contaminated drinking water causes about 505,000 diarrheal deaths a year (WHO). A minority of those deaths happen among people who do have a basic service, so we count about 450,000 for the group above.
  3. People per death. 703 million ÷ 450,000 ≈ 1,560. We round down to about one death a year for every 1,500 people without safe water.
  4. Cost per life saved = cost per person per year × 1,500. DR Congo: $0.64 × 1,500 ≈ $1,000. Papua New Guinea: $0.30 × 1,500 ≈ $450. Cambodia: $0.35 × 1,500 ≈ $500 (using the lower count of 200 students per school).

How to read it. The estimate assumes people drink the treated water every day. If a household uses it only part of the time, the cost per life saved goes up. In DR Congo it is conservative: conflict, hunger and other illness raise death rates in Goma well above the global average, so fewer than 1,500 people are likely needed to prevent one death. In rural Papua New Guinea and Cambodia, where poverty is the main risk and there is no war or major epidemic, it is a reasonable middle estimate. A school system protects students during the school day, so the Cambodia figure is a rough guide. Sources: WHO drinking-water fact sheet; WHO/UNICEF Joint Monitoring Programme. Browse every project →

Solutions by problem

Every major solution, not just the ones we use, for the two water problems we work on. The right tool depends entirely on the water.

See All Details →
See All Details →

Swipe the table sideways to see every column →

SolutionHow it worksBest forTargetsApprox. costMaintenance / lifeKey tradeoff
BoilingHeat to a rolling boil (1 min; 3 at altitude).Emergency stopgapBacteria, viruses, protozoa~$75+/person/yr in fuelNone; not durableFuel & smoke; no residual
ChlorinationFree chlorine disinfects and leaves a residual.Stored/piped water with a supply chainBacteria (6-log), virus (4-log); weak on Crypto~$0.66/person/yr; a $1.50 bleach bottle treats ~20,000 galDosing checks; gear 5–10 yrTaste; fails in turbid water
SODIS (solar)Clear PET bottles in the sun 6 h (2 days if cloudy).High-sun regions, small volumesBacteria, protozoa; weaker on virus~$0.30/person/yrSwap bottles every 6–12 moWeather-dependent; ≤2 L
Ceramic filtersWater percolates fired clay (often silver-lined).Household, locally made99% bacteria, >99% protozoa; limited virus~$3–7/person/yr ($15–50/unit)Scrub; replace 2–5 yrSlow flow; breakable
Biosand / slow sandA living bio-layer plus sand traps & digests pathogens.Durable household/community use>98% bacteria, >99% protozoa, 85–90% virus~$100/unit; ~$0.26–4/person/yr10+ yr; swirl-and-dump top sandHeavy; ~4-week ripening period
Hollow-fiber UF (Sawyer)Gravity 0.1µm membrane; no power.Dispersed households (our DRC default)Bacteria (7-log), protozoa (6-log); limited virus<$0.40 per 1,000 gal; ~$0.11–1.46/person/yrSyringe backflush; 100,000+ gal lifeWon’t remove viruses or chemicals
Powered / community UFPressurized hollow-fiber at kiosk or school scale.With power + operator (our PH UF+UV)Bacteria, protozoa; viruses only with a validated ~0.01µm membrane~$700 school unit (200–300 students); ~$0.25–0.35/person/yrMembrane 5–15 yr; needs powerCapital cost + skilled upkeep
Reverse osmosisHigh-pressure membrane rejects salts & metals.Only when dissolved contaminants demand itSalts, arsenic, fluoride, virusesHigh capex + energy/opexMembranes; pre-treatmentWastes water; strips minerals; costly
UV disinfectionA UV-C lamp damages microbial DNA.Clear water + power, point-of-entryBacteria, viruses, protozoa~$7–69/person/yrUnit 5–10 yr; lamp replaced every 1–3 years depending on the model; clean the sleeveNeeds power; no residual
Coagulation / flocculationAlum or PAC clumps fine particles to settle out.Pre-treating turbid surface waterTurbidity; some pathogens & arsenicLow chemical costDosing & sludge removalA pre-treatment, not disinfection
Rainwater harvestingRoof catchment + first-flush + storage (a source).High-rainfall regionsLow-contaminant source water$2,000–5,000 (storage is >60% of cost)Clean gutters/tank; 10–20 yrSeasonal; storage dominates cost
Protected wells / boreholesA sealed wellhead & apron (a source).Areas with good groundwaterReduces fecal ingressShallow $500–2,000; borehole $15k–40kPump upkeep; apron repairMay carry geogenic arsenic/fluoride
Arsenic removalIron-based adsorption (SONO), coag-filtration, or RO.Arsenic groundwater (WHO 10µg/L)Arsenic (III & V)Iron is cheap; a SONO filter ran 15 yr at 600 ppbReplace the iron matrixNeeds pH 6–7; As(III) needs oxidation
Fluoride removalBone char, activated alumina, or the Nalgonda technique.High-fluoride belts (WHO 1.5 mg/L)FluorideMedia + regeneration costPeriodic media swapCapacity limits; safe disposal

Our field principle: low cost protects more people. 99% safe water for 1,000 beats 100% safe water for 50. Pathogens are the default threat (match a filter, chlorine, or UV to a water test); arsenic and fluoride need contaminant-specific media or RO. Figures are approximate, from CWH field assessments cross-checked with WHO, US EPA, and CAWST.

Swipe the table sideways to see every column →

SolutionHow it worksBest forTargetsApprox. costMaintenance / lifeKey tradeoff
Activated carbon / biocharPorous carbon adsorbs mercury (and cyanide); biochar is community-makeable.Mercury in drinking waterMercury, cyanideLow; can be produced locallyReplace spent mediaMethylmercury biomagnifies (the real danger)
Bone charHydroxyapatite chemically binds lead & cadmium (also fluoride).Lead/cadmium in drinking waterLead, cadmium, fluorideCheap; burn bone locallyReplace when saturatedFinite capacity; safe disposal
Iron-based adsorptionIron hydroxides bind arsenic (and some Pb/Cd).Arsenic-laden waterArsenicIron is cheap and abundantReplace mediaNeeds pH ~6–7
Raise pH (dolomite / limestone)Carbonate raises pH so metals precipitate; adds Ca/Mg.Acidic, metal-rich waterZinc & general metals; acidityLocally abundant (e.g. Ghana)Replenish mediaOnly partial for Pb/Cd/Hg
Ca/Mg ‘safening agent’Dietary calcium/magnesium displaces toxic metals in the body.Reducing harm at a given exposureLowers metal toxicityCheap mineral pelletsSlow-dissolvingMitigates harm; doesn’t remove metal
Mercury retorts (source)Condense & recapture mercury vapor when burning amalgam.Artisanal gold minersReduces mercury releasedCheap per deviceAdoption-dependentOccupational protection only
Gravity-borax (source)Borax + gravity capture gold without any mercury.Artisanal & small-scale miningEliminates mercury useCheaper; recovers more goldTrainingNeeds local leadership / enforcement
AMD neutralization + wetlandsLimestone raises pH; constructed wetlands passively treat.Acid mine drainage (pH 2.5–4)Acidity, dissolved metalsLow opex; wetlands need landManage metal sludgeGenerates sludge to dispose of
Cyanide degradation / green chemNatural degradation, or replace cyanide leaching entirely.Gold extraction (our Baguio R&D)CyanideR&D-stageProcess changeSubstitute still being proven
PhytoremediationHyperaccumulator plants pull metals into harvestable biomass.Soils & sediments, slow cleanupMetals (slowly)Low costHarvest & dispose safelySlow; don’t route biomass into feed
Modified SONO filter (our design)One community-built unit: sand+iron+charcoal (As, Hg) + dolomite (pH, Ca/Mg) + bone char (Pb/Cd) + biochar (Hg, CN).Multi-metal mining pollutionArsenic, mercury, lead, cadmium, zincMostly local materialsNeeds before/after testing + lifespan est.A CWH integrated design, in development

Toxicity roughly follows mercury > arsenic > lead > cadmium > zinc, and acidic mine water makes every metal more soluble. Best solved at the source (mercury-free mining, AMD neutralization) and backstopped with selective adsorption media, which our Modified SONO filter combines into one community-built unit. Spent media and metal sludge must be disposed of safely.

Technical briefs

Full write-ups on the solutions we know best: the chemistry, the design choices, and the numbers. We’re publishing these one at a time.

Want one early? Request a brief →

We publish our mistakes

Honesty about failure is how engineering improves, and how you know you can trust us.

What failed
We broke a UV lamp on install

During commissioning of our first Philippine village system, we broke a UV lamp. It cost time and money, and we rewrote our install checklist because of it. Every system since benefits from that mistake.

What failed (sector-wide)
Free systems get undervalued

Across the sector, water systems given away entirely tend not to last, because people don’t value what they get for free. We now build in a small community contribution, which funds local operators and creates ownership.

What worked
Simple beats sophisticated

In the DRC we deliberately avoided chlorine and complex builds. Gravity membrane filters remove the bacteria that cause cholera and typhoid without consumables or supply chains. The simplest design that solves the real problem wins.

A full field-notes blog is on the way. Subscribe below to get each new write-up.

External resources

We’re a learning center, and we don’t only point to our own work. When someone else explains it better, we link to them.

Want to go deeper?

Want a full technical brief, a techno-economic write-up, or a capacity statement for due diligence? Just ask, or lend your own expertise.

Request a Technical Brief → Volunteer Your Expertise